Disclosed herein are exemplary embodiments of methods, apparatus, and systems for performing content-adaptive deblocking to improve the visual quality of video images compressed using block-based motion-predictive video coding. For instance, in certain embodiments of the disclosed technology, edge information is obtained using global orientation energy edge detection (“OEED”) techniques on an initially deblocked image. OEED detection can provide a robust partition of local directional features (“LDFs”). For a local directional feature detected in the partition, a directional deblocking filter having an orientation corresponding to the orientation of the LDF can be used. The selected filter can have a filter orientation and activation thresholds that better preserve image details while reducing blocking artifacts. In certain embodiments, for a consecutive non-LDF region, extra smoothing can be imposed to suppress the visually severe blocking artifacts.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The method of claim 1, wherein the reconstructing, the buffering, the applying, the determining, the selecting, and the selectively applying are performed in a motion compensation loop of the video decoder.
This invention relates to video decoding, specifically improving motion compensation in a video decoder. The problem addressed is the computational inefficiency and potential quality degradation in traditional motion compensation processes, which can lead to artifacts and increased processing time. The method involves reconstructing a reference frame from encoded video data, buffering the reconstructed frame, and applying a motion compensation filter to the buffered frame. The filter parameters are determined based on the motion characteristics of the video content, and a filter is selected from a set of available filters. The selected filter is then applied to the buffered frame to produce a filtered reference frame. This filtered reference frame is used in the motion compensation loop of the video decoder to improve the accuracy of motion prediction and reduce artifacts. The motion compensation loop processes the filtered reference frame to generate a predicted frame, which is then combined with residual data to reconstruct the final decoded frame. The filtering step ensures that the reference frame is optimized for motion compensation, leading to better visual quality and reduced computational overhead. The method is particularly useful in high-efficiency video coding standards where motion compensation plays a critical role in compression efficiency.
3. The method of claim 1, wherein the selectively applying the selected filter includes evaluating one or more application thresholds such that the selected filter is applied at the edge locations determined based at least in part on the analysis of pixel values of the block.
This invention relates to image processing, specifically to methods for selectively applying filters to edge locations in an image block. The problem addressed is improving image quality by selectively applying filters to edge regions while avoiding unnecessary processing in non-edge areas, which can reduce computational overhead and preserve image details. The method involves analyzing pixel values within an image block to identify edge locations. Once edges are detected, one or more application thresholds are evaluated to determine whether a selected filter should be applied at those edge locations. The thresholds may be based on factors such as edge strength, contrast, or other pixel value characteristics. By dynamically applying filters only where needed, the method enhances edge sharpness while minimizing artifacts in smooth regions. The analysis of pixel values may involve comparing pixel intensities, gradients, or other statistical measures to identify transitions that indicate edges. The selected filter could be a sharpening filter, a noise reduction filter, or another edge-enhancing filter. The thresholds ensure that the filter is applied only when the edge characteristics meet certain criteria, preventing over-processing and maintaining natural image appearance. This approach improves computational efficiency by avoiding unnecessary filtering in non-edge regions and enhances image quality by precisely targeting edge regions for enhancement. The method is particularly useful in applications requiring real-time processing, such as video streaming or high-resolution imaging.
4. The method of claim 1, further comprising repeating the determining, the selecting, and the selectively applying for each of one or more other blocks.
A method for optimizing data processing in a computing system involves analyzing data blocks to identify patterns or characteristics that influence processing efficiency. The method includes determining a processing requirement for a data block, selecting a processing technique based on the determined requirement, and applying the selected technique to the block. The processing techniques may include compression, encryption, or other data manipulation methods tailored to the block's properties. The method further involves repeating this process for additional data blocks, allowing dynamic adaptation of processing strategies across multiple blocks. This approach improves computational efficiency by matching processing techniques to specific data characteristics, reducing unnecessary operations and enhancing performance. The method is particularly useful in systems handling large datasets where uniform processing would be inefficient. By dynamically selecting and applying appropriate techniques, the system can optimize resource usage and processing speed.
9. The method of claim 1, wherein at least one of the two or more candidate filters is a smoothing filter configured to remove discontinuity artifacts while preserving real edges.
A method for image processing involves selecting and applying one or more candidate filters to an image to enhance its quality. The method addresses the problem of improving image clarity by reducing noise and artifacts while maintaining important visual features. Among the candidate filters, at least one is a smoothing filter specifically designed to eliminate discontinuity artifacts—such as jagged edges or abrupt transitions—without blurring genuine edges in the image. This ensures that the processed image retains sharp, natural-looking features while reducing visual distortions. The smoothing filter may be applied alongside other filters, such as noise reduction or sharpening filters, to achieve a balanced enhancement. The method dynamically selects the most effective filters based on the image content, ensuring optimal results for different types of images. This approach improves image quality in applications like medical imaging, surveillance, and digital photography, where clarity and detail preservation are critical.
10. The method of claim 1, wherein the selectively applying the selected filter includes evaluating, using the edge locations determined based at least in part on the analysis of pixel values of the block, whether or not to apply the selected filter.
This invention relates to image processing, specifically to methods for selectively applying filters to image blocks based on edge detection. The problem addressed is the need for efficient and accurate filtering in image processing, particularly in scenarios where different regions of an image require different filtering techniques. Traditional approaches often apply uniform filters across an entire image, which can lead to suboptimal results in areas with distinct features like edges or textures. The method involves analyzing pixel values within a block of an image to determine edge locations. These edge locations are then used to evaluate whether a selected filter should be applied to the block. The evaluation step ensures that filtering is only applied where it is most effective, such as near edges, while avoiding unnecessary processing in regions where filtering may not be beneficial. This selective application improves computational efficiency and enhances image quality by preserving important features. The analysis of pixel values may involve techniques such as gradient calculation, edge detection algorithms, or other methods to identify transitions or discontinuities within the block. The determined edge locations guide the decision-making process for filter application, ensuring that the filtering process is adaptive and context-aware. This approach is particularly useful in applications like video compression, noise reduction, and image enhancement, where precise control over filtering is critical. By dynamically adjusting filter application based on edge information, the method achieves a balance between performance and quality.
13. The one or more computer-readable media of claim 11, wherein the selectively applying the selected filter includes evaluating, using the edge locations determined based at least in part on the analysis of pixel values of the block, whether or not to apply the selected filter.
This invention relates to image processing, specifically to methods for selectively applying filters to image blocks based on edge detection. The problem addressed is improving image quality by dynamically applying filters only where needed, rather than uniformly across an entire image. Traditional approaches often apply filters indiscriminately, which can degrade image quality by blurring edges or introducing artifacts. The invention involves analyzing pixel values within an image block to determine edge locations. These edge locations are used to evaluate whether a selected filter should be applied to the block. The evaluation considers the presence and characteristics of edges to decide if filtering would be beneficial or harmful. For example, if strong edges are detected, the filter may be skipped to preserve sharpness, whereas in smoother regions, the filter may be applied to reduce noise or enhance details. The selective application ensures that filters are only used where they improve image quality without introducing unwanted distortions. The method includes determining edge locations by analyzing pixel value gradients or other edge-detection techniques. The filter selection process may involve comparing the detected edges against predefined criteria, such as edge strength or direction, to decide whether to apply the filter. This adaptive approach optimizes image processing by tailoring filter application to the local image content. The invention can be implemented in software, hardware, or a combination thereof, and is applicable to various image processing applications, including noise reduction, sharpening, and compression.
15. The computer system of claim 14, wherein the reconstructing, the buffering, the applying, the determining, the selecting, and the selectively applying are performed in a motion compensation loop of the video encoder.
This invention relates to video encoding, specifically improving motion compensation in a video encoder. The problem addressed is the computational inefficiency and quality degradation in traditional motion compensation techniques, which often fail to accurately reconstruct reference frames or apply motion vectors effectively. The invention provides a system that reconstructs a reference frame from a bitstream, buffers the reconstructed frame, and applies motion compensation using motion vectors to generate a predicted frame. The system then determines a difference between the predicted frame and an original frame, selects a motion compensation method based on the difference, and selectively applies the method to improve encoding efficiency. The motion compensation loop includes these steps: reconstructing the reference frame, buffering it, applying motion compensation, determining the difference, selecting the appropriate method, and selectively applying the method. This iterative process enhances prediction accuracy and reduces computational overhead, leading to better video compression and quality. The invention is particularly useful in real-time video encoding applications where efficiency and quality are critical.
16. The computer system of claim 14, further comprising repeating the determining, the selecting, and the selectively applying for each of one or more other blocks.
A computer system is designed to optimize data processing by dynamically adjusting operations based on performance metrics. The system monitors the execution of data blocks within a processing pipeline, measuring performance characteristics such as latency, throughput, or resource utilization. Based on these measurements, the system selects an optimization strategy from a predefined set of options. The selected strategy is then applied to the data block to improve efficiency. This process is repeated iteratively for multiple data blocks within the pipeline, allowing the system to continuously adapt to changing conditions. The optimization strategies may include techniques such as load balancing, parallel processing, or resource allocation adjustments. By dynamically applying these strategies, the system enhances overall processing efficiency and reduces bottlenecks. The approach is particularly useful in high-performance computing environments where real-time adjustments are critical for maintaining performance. The system ensures that each data block is processed optimally, leading to improved throughput and reduced latency in data-intensive applications.
18. The computer system of claim 14, wherein at least one of the two or more candidate filters is a smoothing filter configured to remove discontinuity artifacts while preserving real edges.
This invention relates to computer systems for image or signal processing, specifically addressing the challenge of removing discontinuity artifacts while preserving real edges in processed data. The system includes a filtering mechanism that applies two or more candidate filters to an input signal or image. At least one of these filters is a smoothing filter designed to reduce or eliminate discontinuity artifacts, such as noise or abrupt transitions, without blurring or distorting genuine edges in the data. The system may also include additional filters, such as edge-preserving filters, to further enhance the quality of the processed output. The filtering process may involve selecting the most appropriate filter or combining multiple filters to achieve optimal results. The invention is particularly useful in applications where maintaining sharp edges is critical, such as medical imaging, computer vision, or signal processing for audio or video data. The system dynamically adapts to the input data, ensuring that artifacts are minimized while real features remain intact.
19. The computer system of claim 14, wherein the selectively applying the selected filter includes evaluating, using the edge locations determined based at least in part on the analysis of pixel values of the block, whether or not to apply the selected filter.
This invention relates to image processing systems that enhance image quality by selectively applying filters to image blocks based on edge detection. The problem addressed is the need for efficient and accurate edge-aware filtering to improve image sharpness and reduce artifacts without excessive computational overhead. The system analyzes pixel values within an image block to determine edge locations, which are regions where significant pixel intensity changes occur. These edge locations are used to decide whether to apply a selected filter to the block. The filtering process is adaptive, ensuring that filters are only applied where they are most effective, such as near edges, while avoiding unnecessary processing in smooth regions. This selective application improves computational efficiency and enhances image quality by preserving edge details while smoothing non-edge areas. The system may use various filtering techniques, such as deblurring or noise reduction, depending on the detected edge characteristics. By dynamically adjusting filter application based on edge analysis, the invention optimizes image processing performance and accuracy.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
March 1, 2021
December 13, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.